Generalized Hidden Parameter MDPs Transferable Model-based RL in a Handful of Trials

Abstract

There is broad interest in creating RL agents that can solve many (related) tasks and adapt to new tasks and environments after initial training. Model-based RL leverages learned surrogate models that describe dynamics and rewards of individual tasks, such that planning in a good surrogate can lead to good control of the true system. Rather than solving each task individually from scratch, hierarchical models can exploit the fact that tasks are often related by (unobserved) causal factors of variation in order to achieve efficient generalization, as in learning how the mass of an item affects the force required to lift it can generalize to previously unobserved masses. We propose Generalized Hidden Parameter MDPs (GHP-MDPs) that describe a family of MDPs where both dynamics and reward can change as a function of hidden parameters that vary across tasks. The GHP-MDP augments model-based RL with latent variables that capture these hidden parameters, facilitating transfer across tasks. We also explore a variant of the model that incorporates explicit latent structure mirroring the causal factors of variation across tasks (for instance: agent properties, environmental factors, and goals). We experimentally demonstrate state-of-the-art performance and sample-efficiency on a new challenging MuJoCo task using reward and dynamics latent spaces, while beating a previous state-of-the-art baseline with > 10× less data. Using test-time inference of the latent variables, our approach generalizes in a single episode to novel combinations of dynamics and reward, and to novel rewards.

Towards Continual Reinforcement Learning: A Review and Perspectives

“…This has a very practical benefit as it limits the length of applicable interaction history. Current work, also typically assumes a stationary task state distribution for meta-learning (Doshi-Velez and Konidaris, 2013;Wang et al, 2016;Zintgraf et al, 2018;Rakelly et al, 2019;Zintgraf et al, 2019;Humplik et al, 2019;Fakoor et al, 2019;Perez et al, 2020). However, this framework has also has been readily applicable to more challenging multi-agent learning settings (Da Silva et al, 2006;Amato et al, 2013;Marinescu et al, 2017;Vezhnevets et al, 2019).…”